Software workload estimation steps in seven scenarios

Scenario 1: workload estimation before contract

Scenario Description:

(1) No cmmi2 implementation
(2) If the contract is not signed, you need to quote the customer
(3) There are customers' summary requirements and similar project data for reference.
(4) You need to estimate the total workload of the entire project to estimate the total cost and quote the customer

Estimation steps:

(1) Search for similar historical projects, perform project analogy analysis, and estimate the total workload of the project based on experience;
(2) Conduct WBS decomposition to break down the tasks of the entire project as far as possible;
(3) Use empirical methods to estimate the workload of each type of activities in WBS based on data from similar projects;
(4) summarize the total workload of the project;
(5) confirm and analyze the results in step (1), and determine the estimation results based on the analysis results.

Scenario 2: Experience estimation based on detailed requirements

Scenario Description:

(1) only detailed requirements, no historical data

Estimation steps:

(1) WBS decomposition: splits tasks into the granularity that a person or a small team can execute. WBS should identify all deliverables,Project Management


Activities, engineering activities, etc.
(2) Use the empirical method to estimate the workload of each activity;
(3) Summary: the workload of each stage and the total workload of the project.

Other Instructions:

In this scenario, only the empirical method is used to verify the results, and it is difficult to judge the rationality of the results.

Scenario 3: overall estimation by encoding

Scenario Description:

(1) historical data of similar projects
(2) productivity data with coding activities
(3) have detailed requirements
(4) implemented cmmi2 level, but did not accumulate the workload distribution data of historical projects

Estimation steps:

(1) Product decomposition: The system is divided into subsystems and subsystems are divided into modules;
(2) WBS decomposition: splits tasks into the granularity that a person or a small team can execute. WBS should identify all deliverables, project management activities, and engineering activities.
(3) Establish the ing relationship between activities in The WBS decomposition and product elements, and identify which activities in the WBS can be estimated using the model method;
(4) estimate the scale of product elements.Code


Line method or function point method, and estimate the complexity and Reuse Rate of each product element;
(5) Calculate the coding workload of each product element by model based on the productivity data in the historical coding stage and the scale estimation, complexity, and reuse rate of product elements;
(6) The workload of other activities can be estimated based on the data of similar projects in history and the experience of the estimator.
(7) Summary: the workload of each stage and the total workload of the project.

Other Instructions:

In this scenario, the empirical method and model method are used in combination. These two methods complement each other, rather than mutual validation.

Scenario 4: WBS-based Estimation

Scenario Description:

(1) historical data of similar projects
(2) productivity data (including management workload) for similar projects throughout the lifecycle)
(3) have detailed requirements
(4) implemented cmmi2 level, but did not accumulate the workload distribution data of historical projects

Estimation steps:

(1) Product decomposition: The system is divided into subsystems and subsystems are divided into modules;
(2) The code line method or function point method can be used to estimate the scale of product elements;
(3) accumulate the total scale of the entire product, and estimate the overall complexity and Reuse Rate of the product;
(4) calculate the total development workload based on the productivity data of similar projects throughout the life cycle and the total scale, complexity, and reuse rate of products;
(5) WBS decomposition: splits tasks into the granularity that a person or a small team can execute. WBS should identify all deliverables, project management activities, and engineering activities.
(6) the workload of all activities can be estimated based on the data of similar projects in history and the experience of the estimator.
(7) Summary: the workload of each stage and the total workload of the project.
(8) compare with the workload obtained in step (4). If the deviation is not large, the result of step (7) shall prevail. If the deviation is large, the possible causes should be carefully analyzed as follows:
Productivity Data of similar projects is not suitable for other projects;
The granularity of WBS decomposition is not detailed enough;
Estimating experts' experience is not suitable for the project;
The estimation of specific tasks is unreasonable;
Adjust the estimation result to make it more reasonable.

Other Instructions:

In this scenario, there are two or more results for the total workload of the project. These results can be mutually verified to find out the irrationality in the estimation process, which is more reasonable.

Scenario 5: 3D Validation Based on WBS Estimation

Scenario Description:

(1) historical data of similar projects
(2) productivity data (including management workload) for similar projects throughout the lifecycle)
(3) have detailed requirements
(4) implemented cmmi3-level workload distribution data for historical projects (phase distribution and job type distribution)

Estimation steps:

(1) Product decomposition: The system is divided into subsystems and subsystems are divided into modules;
(2) The code line method or function point method can be used to estimate the scale of product elements;
(3) accumulate the total scale of the entire product, and estimate the overall complexity and Reuse Rate of the product;
(4) calculate the total development workload based on the productivity data of similar projects throughout the life cycle and the total scale, complexity, and reuse rate of products;
(5) Based on the workload distribution data of historical projects and the total workload estimated in step (4), the calculation is as follows:

• Workload at each stage
• Workload for each type of work

(6) WBS decomposition: splits tasks into the granularity that a person or a small team can execute. WBS should identify all deliverables, project management activities, and engineering activities.
(7) the workload of all activities can be estimated based on the data of similar projects in history and the experience of the estimator.
(8) summarized: the workload of each stage, the workload of each job, and the total workload of the project.
(9) compare with the workload obtained in steps (4) and (5). If the deviation is not large, the result of step (7) shall prevail. If the deviation is large, the possible causes should be carefully analyzed as follows:
Productivity Data of similar projects is not suitable for other projects;
The granularity of WBS decomposition is not detailed enough;
Estimating experts' experience is not suitable for the project;
The estimation of specific tasks is unreasonable;
Adjust the estimation result to make it more reasonable.

Other Instructions:

In this scenario, there are two or more results for the total workload of the project, and the two methods are used to get the workload of each stage, each job, and the total workload of the project, these results can be verified from the above three dimensions to find that the estimation process is unreasonable, which is more reasonable.

Scenario 6: four-dimensional Validation Based on WBS Estimation

Scenario Description:

(1) historical data of similar projects
(2) productivity data with code activities for similar projects (excluding management workload)
(3) have detailed requirements
(4) Implementation of cmmi3 level, workload distribution data of historical projects (phase distribution, type distribution, phase type distribution)
(5) The project adopts the waterfall model.

Estimation steps:

(1) Product decomposition: The system is divided into subsystems and subsystems are divided into modules;
(2) estimate the scale of product elements. You can use the code line method or the function point method to estimate the complexity and Reuse Rate of each product element;
(3) Calculate the coding workload of each product element by model based on the productivity data of coding activities of similar projects and the scale, complexity, and reuse rate of product elements;
(4) Calculate the workload based on the workload distribution of historical projects and the estimated coding workload in step (3:
I) calculate the total workload of the encoding phase based on the coding workload of historical projects as a percentage of the workload in the encoding phase and the coding workload calculated in part (3;
Ii) Calculate the workload of each type of work in the coding phase based on the workload distribution percentage of each type of work in the coding phase of the historical project;
Iii) calculate the total workload of other phases based on the workload of other phases of the historical project and the workload ratio of the coding phase;
Iv) Calculate the workload of each type of work in other stages of the history project based on the workload distribution percentage of each type of work and step III;
(5) WBS decomposition: splits tasks into the granularity that a person or a small team can execute. WBS should identify all deliverables, project management activities, and engineering activities.
(6) the workload of all activities can be estimated based on the data of similar projects in history and the experience of the estimator.
(7) Summary: the workload of each type of work in each stage, the workload of each stage, the workload of each type of work, and the total workload of the project.
(8) compare with the workload obtained in step (4). If the deviation is not large, the result of step (6) shall prevail. If the deviation is large, the possible causes should be carefully analyzed as follows:
Productivity Data of similar projects is not suitable for other projects;
The granularity of WBS decomposition is not detailed enough;
Estimating experts' experience is not suitable for the project;
The estimation of specific tasks is unreasonable;
Adjust the estimation result to make it more reasonable.

Other Instructions:

In this scenario, there are two or more results for the total workload of the project, the two methods are used to obtain the workload of each stage, the workload of each job, the workload of each job in each stage, and the total workload of the project, these results can be verified from the above four dimensions to find that the estimation process is unreasonable, which is more reasonable.

Scenario 7: workload Estimation for demand changes

Scenario Description:

(1) requirements for changes
(2) The project has reached the coding stage.
(3) coding productivity for this project

Estimation steps:

(1) analyze the impact scope of demand changes
(2) Conduct the WBS decomposition of this change
(3) estimate the size, complexity, and other attributes of code changes caused by changes
(4) Workload estimation using the Model Method Based on the coding productivity and estimated scale of the project
(5) experience estimation of other activities in WBS Decomposition
(6) summarize all the workload and estimate the workload of the change.

 

Article Source

Http://space.itpub.net/14780828/viewspace-621755